CN110909827A - Noise reduction method suitable for fan blade sound signals - Google Patents
Noise reduction method suitable for fan blade sound signals Download PDFInfo
- Publication number
- CN110909827A CN110909827A CN201911301369.5A CN201911301369A CN110909827A CN 110909827 A CN110909827 A CN 110909827A CN 201911301369 A CN201911301369 A CN 201911301369A CN 110909827 A CN110909827 A CN 110909827A
- Authority
- CN
- China
- Prior art keywords
- noise
- frame
- signal
- wind
- signals
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000005236 sound signal Effects 0.000 title claims abstract description 27
- 238000000034 method Methods 0.000 title claims abstract description 22
- 238000012545 processing Methods 0.000 claims abstract description 13
- 238000007781 pre-processing Methods 0.000 claims abstract 3
- 238000009432 framing Methods 0.000 claims description 5
- 238000012935 Averaging Methods 0.000 claims description 2
- 230000003595 spectral effect Effects 0.000 claims description 2
- 238000003064 k means clustering Methods 0.000 claims 2
- 238000003745 diagnosis Methods 0.000 description 4
- 238000011160 research Methods 0.000 description 3
- 238000013144 data compression Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000037433 frameshift Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 239000013598 vector Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F03—MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
- F03D—WIND MOTORS
- F03D17/00—Monitoring or testing of wind motors, e.g. diagnostics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/14—Fourier, Walsh or analogous domain transformations, e.g. Laplace, Hilbert, Karhunen-Loeve, transforms
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L21/00—Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
- G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
- G10L21/0208—Noise filtering
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/03—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters
- G10L25/24—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters the extracted parameters being the cepstrum
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F05—INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
- F05B—INDEXING SCHEME RELATING TO WIND, SPRING, WEIGHT, INERTIA OR LIKE MOTORS, TO MACHINES OR ENGINES FOR LIQUIDS COVERED BY SUBCLASSES F03B, F03D AND F03G
- F05B2260/00—Function
- F05B2260/80—Diagnostics
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02B—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO BUILDINGS, e.g. HOUSING, HOUSE APPLIANCES OR RELATED END-USER APPLICATIONS
- Y02B10/00—Integration of renewable energy sources in buildings
- Y02B10/30—Wind power
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Computational Mathematics (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Acoustics & Sound (AREA)
- Computational Linguistics (AREA)
- Signal Processing (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Bioinformatics & Computational Biology (AREA)
- Sustainable Development (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Chemical & Material Sciences (AREA)
- Combustion & Propulsion (AREA)
- Mechanical Engineering (AREA)
- Quality & Reliability (AREA)
- Soundproofing, Sound Blocking, And Sound Damping (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Sustainable Energy (AREA)
- Probability & Statistics with Applications (AREA)
- Evolutionary Biology (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
- Artificial Intelligence (AREA)
Abstract
The invention provides a noise reduction method suitable for a fan blade sound signal, which comprises the following steps: firstly, establishing wind noise databases under different wind speed scales, and preprocessing the sound signals of the blades to be denoised through the characteristics of left and right sound channels; then taking out a wind noise signal at the same wind speed as the signal to be subjected to noise reduction as a reference signal, and finding out a frame containing noise and a frame not containing noise of the signal to be subjected to noise reduction based on a Pearson correlation coefficient and a K _ means algorithm; and finally, processing the two types of frame signals in different modes, and performing frame restoration to obtain pure sound signals of the fan blade. The invention can realize noise reduction pretreatment on sound signals in a wind field and filter the interference of wind noise existing for a long time.
Description
Technical Field
The invention relates to a noise reduction method suitable for a fan blade sound signal, and designs a flow method capable of eliminating wind noise and other random noises aiming at the problem of interference of the wind noise and other random noises when the fan sound is collected in a wind field, belonging to the technical field of sound signal noise reduction.
Background
The blade fault detection of the wind power plant is always a problem which cannot be ignored, if the fault can not be diagnosed at the initial stage of the blade fault, along with the increase of the operation time, the fault degree can be further deepened, the maintenance cost is increased, the capture efficiency of wind energy is reduced, and the service life of the blade can be shortened in serious cases.
Numerous scholars at home and abroad make a great deal of research in the group navigation diagnosis of the blades, the fault diagnosis of the blades is rarely researched based on the sound signals of the fan blades, the wind noise is mainly influenced when the sound signals of the fan are collected in a wind field, the fault diagnosis based on the sound signals is greatly interfered, and the method is also a key link for restricting the research, so that the wind noise is eliminated, and the pure sound signals of the fan blades have profound significance in the fault diagnosis of the blades.
Aiming at the noise reduction problem of the fan, the common method is to realize noise reduction based on algorithms such as a filter, wavelet transformation, EMD decomposition and the like, and different from the algorithms, the invention realizes the purpose of noise reduction by building a wind noise signal database under different wind speed scales and re-fitting a noise-containing frame signal from the angle of framing.
Disclosure of Invention
The invention aims to research a noise reduction method suitable for a fan blade sound signal, aiming at a wind field, firstly establishing a wind noise signal database under different wind speed scales, framing the sound signal to be denoised and extracting the Mel Frequency Cepstrum Coefficient (MFCC) of each frame; then searching a noisy frame and a non-noisy frame according to the correlation of the wind noise signal, and carrying out different processing on the two types of frame signals; and finally, reconstructing each frame signal to obtain a pure sound signal of the fan blade, and realizing the noise reduction processing of the sound signal.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flow chart of a noise reduction method for a fan blade acoustic signal according to the present disclosure;
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
As shown in fig. 1, a flow chart of a noise reduction method for a fan blade sound signal according to the present invention includes the steps of:
step 1, wind noise in a wind field is random and does not exist all the time, so that wind noise signals under different wind speed scales are collected by a single sound track collecting device in a field far away from a fan aiming at a certain wind field, and a database of wind noise information of the wind field is established, wherein the interval of the wind speed scales is 1 m/s;
step 2, aiming at the blade sound signal x needing noise reduction processing1(t) collecting by using a dual-channel sound collecting device, respectively depicting waveforms of signals to be denoised of a left channel and a right channel on a time domain, and for a waveform amplitude value of a certain time point, taking the minimum value of the waveform amplitude values of the left channel and the right channel corresponding to the time point, and according to the principle, obtaining a blade sound signal x preprocessed based on the left channel and the right channel2(t)；
Step 3, taking out a reference wind noise signal z (t) with the same wind speed as the signal to be noise-reduced from a wind noise database, and aiming at z (t) and x respectively2(t) the same framing is performed, whereinx2(t) is divided into n1Frame, z (t) is divided into n2Frame, framing formula is as follows:
fn＝(N-wlen+inc)/inc (1)
overlap＝wlen-inc (2)
where N is the length of the audio signal, wlen is the set frame length, inc is the set frame shift, typically around 1/4 of the frame length, overlap is the frame overlap, fn is the number of frames into which the signal is divided;
MFCC is a data compression technique, which can use a group of vectors composed of 12-16 values to represent the characteristic situation of a frame signal, extract z (t) and x2(t) MFCC coefficients per frame, i.e., MFCC (z, n)2)、mfcc(x2,n1)；
Step 4. for x2(t) mfcc (x) of signals per frame2K) respectively with z (t) and then averaging to obtain x2(t) the correlation of each frame signal with the reference wind noise R (k), where k is 1-n1Q is 1-n2The formula involved is as follows:
equation (3) shows that the process Cov (x, y) for finding the Pearson correlation coefficient of the two variables x, y is the covariance of the variables x, y, σxAnd σyVariance of x and y respectively;
aiming at a data set formed by R (K), clustering is carried out by adopting a K _ means method, the clustering number is 2, in a clustering result, R (K) in a class with a smaller numerical value corresponds to a non-noise frame without noise, and R (K) in a class with a larger numerical value corresponds to a noise frame with noise;
step 5, reserving the non-noise frames without any processing, and regarding the noise-containing frames, taking any frame of wind noise frames to perform Fourier transform (FFT), subtracting the noise-containing frames from the wind noise frames in the frequency domain, and re-depicting spectral lines in the frequency domain to obtain the processing results of the noise-containing frames;
step 6, carrying out frame reduction processing on the processed noisy frame and the non-noisy frame, and taking out x2(t) 1-overlap data points for frame 1 and frames 2 through n1And recombining a row of data by overlap + 1-wlen data points in each frame, wherein the row of data is the noise-reduced blade sound signal.
The principle and the implementation mode of the invention are explained by applying specific embodiments in the invention, and the description of the embodiments is only used for helping to understand the method and the core idea of the invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.
Claims (6)
1. A method of noise reduction for a fan blade acoustic signal, the method comprising:
step 1, establishing wind noise sound signal databases under different scales;
step 2, preprocessing blade sound signals needing noise reduction based on left and right sound channels;
step 3, taking out wind noise signals at the same wind speed as the signals to be denoised, then carrying out the same framing processing on the two signals, and extracting the MFCC coefficients of the two signals;
step 4, obtaining a correlation coefficient R (K) of each frame data of the signal to be denoised and a wind noise frame, and identifying a frame signal without noise and a frame signal with noise by using a K _ means clustering method;
step 5, reserving the non-noise frame signal, and performing difference processing on the noise-containing frame signal and the wind noise frame signal in a frequency domain;
and 6, carrying out frame reduction on the non-noise frame signal and the noise frame signal so as to obtain a pure sound signal of the fan blade, and realizing noise reduction processing of sound.
2. The method of claim 1, step 2, wherein the left and right channel based preprocessing is performedThe blade sound signal to be noise reduced is characterized by: the acoustic factor sensor for collecting signals is a dual-channel sensor, waveforms of signals to be denoised of a left channel and a right channel are respectively drawn on a time domain, for a waveform amplitude value of a certain time point, the minimum value of the waveform amplitude values of the left channel and the right channel corresponding to the time point is taken, and according to the principle, a blade sound signal x preprocessed based on the left channel and the right channel is obtained2(t)。
3. The correlation coefficient R (k) of each frame data of the signal to be denoised and the wind noise frame in step 4 of claim 1 is characterized by: for x2(t) mfcc (x) of signals per frame2K) respectively with z (t) and then averaging to obtain x2(t) the correlation value R (k) of each frame signal and the reference wind noise, wherein the formula of R (k) is as follows:
in which k is 1-n1Q is 1-n2。
4. The method of claim 1 and step 4 for identifying the noisy frames and the non-noisy frames by using the K means clustering method is characterized in that: and setting the number of clusters to be 2 for the data set formed by R (k), wherein in the clustering result, R (k) in the class with smaller value corresponds to a non-noise frame without noise, and R (k) in the class with larger value corresponds to a noise frame with noise.
5. A method for processing noisy frames and non-noisy frames as claimed in claim 1 and step 5 is characterized by: and reserving the non-noise frame without any processing, and for the noise-containing frame, taking a wind noise frame to perform Fourier transform, subtracting the noise-containing frame from the wind noise frame in the frequency domain, and re-describing spectral lines in the frequency domain to obtain a processing result of the noise-containing frame.
6. The frame restoration process according to claim 1 or 6 is characterized by: take out x2(t) 1-overlap data points for frame 1 and frames 2 through n1And recombining a row of data by overlap + 1-wlen data points in each frame, wherein the row of data is the noise-reduced blade sound signal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911301369.5A CN110909827A (en) | 2019-12-17 | 2019-12-17 | Noise reduction method suitable for fan blade sound signals |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911301369.5A CN110909827A (en) | 2019-12-17 | 2019-12-17 | Noise reduction method suitable for fan blade sound signals |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110909827A true CN110909827A (en) | 2020-03-24 |
Family
ID=69826042
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911301369.5A Pending CN110909827A (en) | 2019-12-17 | 2019-12-17 | Noise reduction method suitable for fan blade sound signals |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110909827A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113049252A (en) * | 2021-03-25 | 2021-06-29 | 成都天佑路航轨道交通科技有限公司 | Fault detection method for train bearing box |
| CN114530163A (en) * | 2021-12-31 | 2022-05-24 | 安徽云磬科技产业发展有限公司 | Method and system for recognizing life cycle of equipment by adopting voice based on density clustering |
| CN115547356A (en) * | 2022-11-25 | 2022-12-30 | 杭州兆华电子股份有限公司 | Wind noise processing method and system based on abnormal sound detection of unmanned aerial vehicle |
-
2019
- 2019-12-17 CN CN201911301369.5A patent/CN110909827A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113049252A (en) * | 2021-03-25 | 2021-06-29 | 成都天佑路航轨道交通科技有限公司 | Fault detection method for train bearing box |
| CN113049252B (en) * | 2021-03-25 | 2023-04-14 | 成都天佑路航轨道交通科技有限公司 | Fault detection method for train bearing box |
| CN114530163A (en) * | 2021-12-31 | 2022-05-24 | 安徽云磬科技产业发展有限公司 | Method and system for recognizing life cycle of equipment by adopting voice based on density clustering |
| CN115547356A (en) * | 2022-11-25 | 2022-12-30 | 杭州兆华电子股份有限公司 | Wind noise processing method and system based on abnormal sound detection of unmanned aerial vehicle |
| CN115547356B (en) * | 2022-11-25 | 2023-03-10 | 杭州兆华电子股份有限公司 | Wind noise processing method and system based on abnormal sound detection of unmanned aerial vehicle |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109145727A (en) | A kind of bearing fault characteristics extracting method based on VMD parameter optimization | |
| Wang et al. | ia-PNCC: Noise Processing Method for Underwater Target Recognition Convolutional Neural Network. | |
| Jangjit et al. | A new wavelet denoising method for noise threshold | |
| CN101853665A (en) | Method for eliminating noise in voice | |
| CN108922514B (en) | Robust feature extraction method based on low-frequency log spectrum | |
| CN110909827A (en) | Noise reduction method suitable for fan blade sound signals | |
| CN102982808A (en) | Speech denoising device and method based on wavelet transform | |
| Huang et al. | Dccrgan: Deep complex convolution recurrent generator adversarial network for speech enhancement | |
| CN103578466A (en) | Voice and non-voice detection method based on fractional order Fourier transformation | |
| Li | Speech denoising based on improved discrete wavelet packet decomposition | |
| CN112185405B (en) | Bone conduction voice enhancement method based on differential operation and combined dictionary learning | |
| Yu et al. | Sparse time–frequency representation for the transient signal based on low-rank and sparse decomposition | |
| CN115655719A (en) | Bearing vibration signal staged noise reduction method and bearing fault identification method | |
| CN116312561A (en) | Method, system and device for voice print recognition, authentication, noise reduction and voice enhancement of personnel in power dispatching system | |
| CN113571074B (en) | Voice enhancement method and device based on multi-band structure time domain audio frequency separation network | |
| CN116665681A (en) | Thunder identification method based on combined filtering | |
| CN111968627B (en) | Bone conduction voice enhancement method based on joint dictionary learning and sparse representation | |
| TWI749547B (en) | Speech enhancement system based on deep learning | |
| CN114858430A (en) | Mechanical seal acoustic emission signal noise reduction method based on LCD-wavelet new threshold | |
| Tayachi et al. | A New Approach for Detection of Gear Defects using a Discrete Wavelet Transform and Fast Empirical Mode Decomposition | |
| CN113343887A (en) | Multi-sensor mixed fault signal blind separation method based on edge calculation and machine learning | |
| Feng et al. | Voice activity detection based on ensemble empirical mode decomposition and teager kurtosis | |
| CN106997766B (en) | Homomorphic filtering speech enhancement method based on broadband noise | |
| Sudheer Kumar et al. | Noise Reduction in Audio File Using Spectral Gatting and FFT by Python Modules | |
| Fenghua et al. | Speech signal enhancement through wavelet domain MMSE filtering |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |